Query usage statistics from all your AI coding tools in one command - Claude Code, Cursor, GitHub Copilot, Windsurf, Warp AI, and 14+ tools
npm install a2zusage
Query usage statistics from ALL your AI coding tools in one command
Installation •
Supported Tools •
Usage •
Why This Tool •
Contributing
---
```
╭────────────────┬──────────┬───────────────┬─────────────┬─────────────┬───────────────╮
│ Tool │ Status │ Today │ This Week │ This Month │ Total │
├────────────────┼──────────┼───────────────┼─────────────┼─────────────┼───────────────┤
│ Claude Code │ ✓ Active │ 321K tokens │ 1.4M tokens │ 6.1M tokens │ 24.2M tokens │
│ Cursor │ ✓ Active │ - │ - │ - │ 32.5K tokens │
│ Windsurf │ ✓ Active │ 1 req │ 1 req │ 1 req │ 21 reqs │
│ Warp AI │ ✓ Active │ - │ - │ 8.2M tokens │ 167.1M tokens │
│ Gemini CLI │ ✓ Active │ - │ 72K tokens │ 101K tokens │ 48.1M tokens │
│ GitHub Copilot │ ○ N/A │ - │ - │ - │ - │
│ ... │ ... │ ... │ ... │ ... │ ... │
╰────────────────┴──────────┴───────────────┴─────────────┴─────────────┴───────────────╯
`bash`
npx a2zusage@latest
`bash`
npm install -g a2zusage
`bash`
cargo install a2zusage
`bash`
git clone https://github.com/aezizhu/a2zaiusage.git
cd a2zaiusage
cargo build --release
./target/release/a2zusage
a2zusage supports 14+ AI coding tools out of the box:
| Tool | Data Source | What’s Accurate |
|------|-------------|----------------|
| Claude Code | Local JSONL (~/.claude/projects/) | ✅ Exact token counts (input/output + cache tokens when present) |~/.codeium/
| Cursor | SQLite database | ✅ Exact token counts (when present in DB) |
| GitHub Copilot | GitHub API + Local logs | ⚠️ Usage count / requests only (GitHub does not expose reliable token totals here) |
|| Windsurf | Cascade sessions () | ℹ️ Session count only (token data encrypted). Visit windsurf.ai for detailed usage. |~/.gemini/tmp/
| Warp AI | SQLite database | ✅ Total tokens (Warp does not expose a reliable input/output split) |
| Cline / Roo Code | VS Code extension storage | ✅ Exact token counts (when stored by the extension) |
| OpenCode | Local JSON files | ✅ Exact token counts (when present in session/message usage fields) |
| OpenAI Codex | OpenAI Usage API | ✅ Exact token counts (requires API key + org access) |
|| Gemini CLI | Native sessions () | ✅ Exact token counts from native session files (no setup required) |
| Amazon Q Developer | Local logs | ⚠️ Best-effort: logs may not contain token totals |
| Tabnine | Local logs | ⚠️ Partial: uses explicit token fields when present; no invented prompt/context tokens |
| Gemini Code Assist | Google Cloud | ⚠️ Not implemented in this repo yet |
| Sourcegraph Cody | VS Code extension | ⚠️ Token counts only when present; otherwise request_count only |
| Replit Ghostwriter | Web link | 🔗 Link Only |
`bashQuery all detected AI tools
a2zusage
$3
`bash
a2zusage -t claude-code # Only Claude Code
a2zusage -t cursor # Only Cursor
a2zusage -t warp # Only Warp AI
`$3
`bash
a2zusage -f table # Pretty table (default)
a2zusage -f json # JSON output
a2zusage -f csv # CSV output
`$3
`bash
a2zusage doctor # Check paths and configuration
a2zusage list # List all supported tools
a2zusage --help # Show help
a2zusage -v # Verbose mode with data sources
`$3
`json
[
{
"name": "claude-code",
"display_name": "Claude Code",
"status": "active",
"usage": {
"today": { "input_tokens": 306929, "output_tokens": 14196, "request_count": 2709 },
"this_week": { "input_tokens": 1106656, "output_tokens": 278247, "request_count": 4663 },
"this_month": { "input_tokens": 3902283, "output_tokens": 2166653, "request_count": 32956 },
"total": { "input_tokens": 13648429, "output_tokens": 10593772, "request_count": 89580 }
},
"data_source": "/Users/you/.claude/projects"
}
]
`Why This Tool?
$3
We created this tool because we needed a way to identify developers who truly embrace AI-assisted coding (what we call "Vibe Coding").
When hiring, we found that:
- Resume skills don't tell the whole story - Many claim AI proficiency but rarely use it
- Token usage reveals real habits - High usage = deep integration into daily workflow
- Multiple tool usage shows adaptability - The best devs try everything and use what works
We used a2zusage in our hiring process and found it incredibly effective at identifying candidates who are genuinely productive with AI tools. Now we're open-sourcing it for the community.
$3
Ever wondered:
- How many tokens you've used across all AI coding tools?
- Which tool you use most frequently?
- How your usage has changed over time?
a2zusage answers all these questions in seconds.
$3
- 100% Local - All data is read from local files on your machine
- No Network Calls - Unless you explicitly use API-based providers
- No Data Collection - We never see your usage data
- Open Source - Audit the code yourself
Environment Variables
For API-based providers, set these environment variables:
`bash
GitHub Copilot (or use
gh auth login)
export GITHUB_TOKEN=ghp_xxxOpenAI Codex
export OPENAI_API_KEY=sk-xxxAWS (for Amazon Q)
export AWS_PROFILE=default
`Cross-Platform Support
| Platform | Status |
|----------|--------|
| macOS (Intel) | ✅ Supported |
| macOS (Apple Silicon) | ✅ Supported |
| Linux (x64) | ✅ Supported |
| Linux (ARM64) | ✅ Supported |
| Windows (x64) | ✅ Supported |
Contributing
We welcome contributions! Here's how you can help:
$3
Know of an AI coding tool we don't support? We'd love to add it!
1. Fork the repository
2. Add a new provider in
src/providers/
3. Update the provider registry in src/providers/mod.rs
4. Submit a pull request$3
- Better token parsing
- More accurate cost estimation
- Additional data sources
$3
Found a bug or have a suggestion? Open an issue!
Citation
If you use a2zusage in your research, hiring process, or project, please cite:
`bibtex
@software{a2zusage,
author = {aezizhu},
title = {a2zusage: AI Coding Tools Usage Tracker},
url = {https://github.com/aezizhu/a2zaiusage},
year = {2026}
}
`Or simply link to:
https://github.com/aezizhu/a2zaiusage`Apache License 2.0 - see LICENSE for details.
---
Built with ❤️ for the AI-native developer community